Teams at Cantina Labs

Build and productionize large-scale TTS and adjacent speech models end-to-end: model architecture, training, alignment, evaluation, data curation, distributed GPU scaling, latency/cost profiling, tooling, and safety/mitigation for speech systems.

Design, deploy, and scale low-latency inference infrastructure for generative audio models (TTS, ASR, voice conversion). Build high-performance inference engines, Kubernetes-based autoscaling, CI/CD pipelines, observability, and GPU optimization to bridge research and production for streaming and batch workloads.

The role involves designing and developing model evaluation pipelines, user studies for subjective evaluations, and automated dashboards to report results, alongside leading the evaluation team and communicating with other teams to improve model performance.